Many of the studies that have compared the instructional effectiveness of static with dynamic images have not controlled all the moderating variables involved. This problem is present not only in instructional pictures concerning the curricular topics (e.g., science, technology, engineering and mathematics: STEM), but also in those depicting extracurricular tasks (e.g., human movement tasks). When factors such as appeal, media, realism, size, and interaction are not tightly controlled between statics and animations, researchers may often be comparing apples with oranges. In this review, we provide a categorization of these confounding variables and offer some possible solutions to generate more tightly controlled studies. Future research could consider these biases and solutions, in order to design more equivalent visualizations. As a result, more conclusive evidence could be obtained identifying the boundary conditions for when static or dynamic images are more suitable for educational purposes, across both curricular and extracurricular tasks. (C) 2016 Elsevier Ltd. All rights reserved.